writing-plans

Invoke the writing-plans skill with your spec or requirements. It maps out the file structure first, then decomposes the work into tasks with exact file paths, failing tests, minimal implementations, and commit steps saved to docs/superpowers/plans/.

Each step should be one action taking 2-5 minutes: write the failing test, run it to confirm failure, implement minimal code, run tests to confirm pass, then commit. This keeps progress trackable with checkbox syntax.

Yes. After saving the plan, you choose between subagent-driven development, which dispatches a fresh subagent per task with review between tasks, or inline execution with batch checkpoints using the executing-plans skill.

Common failures include placeholders like TODO or 'add error handling', missing exact file paths, steps without actual code, and inconsistent type or method names across tasks. The skill's self-review checklist catches spec gaps, placeholders, and type mismatches.

Split when the spec covers multiple independent subsystems. Each plan should produce working, testable software on its own, so one plan per subsystem keeps scope manageable and independently executable.